Spectrum Sensing Using Multiple Antennas for Spatially and Temporally Correlated Noise Environments

Jitvanichphaibool, K.,Ying-Chang Liang,Yonghong Zeng

DOI: https://doi.org/10.1109/DYSPAN.2010.5457864

2010-01-01

Abstract:This paper is interested in spectrum sensing using multiple antennas under spatially and temporally correlated noise environments. We exploit cyclostationary features of the primary user's signal in terms of cyclic spectral coherence function and the proposed modified cyclic spectral density function, which has less computational complexity. Two types of detectors are proposed: pre-combining and post-combining detectors. For pre-combining method, a blind maximum ratio combining technique is considered. All detectors are designed to handle noise uncertainty and also be effective in both white noise and colored noise scenarios. Numerical results are given to illustrate the performance of all detectors and verify their efficiency against the noise correlation effect. With the use of estimated channels, pre-combining detectors are superior to post-combining detectors, which do not require channel information. It is also shown that the modified cyclic spectral density function achieves comparable performance to the cyclic spectral coherence function.